Securing Canada’s Advantage in Artificial Intelligence

Version Francophone: ASSURER L’AVANTAGE DU CANADA EN MATIèRE D’INTELLIGENCE ARTIFICIELLE_October2024.pdf

I have spent the last 25 years in the Technology sector, working for market-defining organizations like Dell, Microsoft and now Equinix, in various business development, strategy and Sr leadership roles. I can attest firsthand that the disruption and opportunity occurring as a result of Artificial Intelligence (AI) is incomparable to any other compelling moment in the Tech or Business sectors during my 3 decades in this space (words that seem shallow in comparison to the impact at hand).

AI and more specifically Generative AI (GenAI) is transforming the world as we know it, disrupting critical sectors like education, healthcare, media, and politics at an unprecedented pace. OpenAI’s ChatGPT, currently the most popular GenAI platform globally, reached one hundred million users in just two months while it took apps like TikTok 9 months to reach the same adoption, and Instagram 30 months.

According to IDC, the Global Gen-AI market is forecasted to reach $151 billion by 2027 (4x its current size), while AI spend on Infrastructure is forecasted to reach 18.3 billion by the same period. By 2028, its expected that 1 billion AI applications will have been deployed globally while, today close to 35% of all applications incorporate some version of AI.

According to Deloitte’s Q3 GenAI Report published in August 2024, 38% of CEOs cited that GenAI will disrupt their competitive advantage aggravated by the fact that 50% of organizations still run on legacy code. To add, 2 of 3 of companies are increasing their investments in Gen-AI and 58% respondents also reported realizing a diverse range of benefit from the technology.

Widespread Adoption and investment in AI/Gen-AI are indeed foregone conclusions.

Through AI, we have an incredible opportunity to help drive Canada’s economic prosperity and quality of life while being an example for the rest of the world by playing a role in both the creation and consumption of equitable, responsible AI that align to Canadian social values and norms. ?To do so, it will require a shift in mindset from Canadian leaders, strategic investments, and strategic partnerships.

As a Corporate Executive in the Canadian technology sector and a proud resident of our great country, I am eager to see Canadian provincial and federal governments accelerate their efforts to align oversight and policy with private capabilities to cement Canada’s relevancy in the AI space.

This article aims to make a case for some of the changes required to achieve this endeavor by diving into three topics, prefaced by providing a quick overview of the Global AI landscape.

1.?????? The Opportunity for Canada

2.?????? Unpacking the AI Value Chain

3.?????? The Role of Public-Private Partnerships in Canada’s National AI Strategy (with a spotlight on energy distribution)

?“Today close to 35% of all applications incorporate some version of AI”

Preface: AI: A Global Context

To date, Canada's AI policy emphasizes ethical considerations, transparency, and accountability. ?This not unlike the European Union’s (EU) regulatory approach with its proposed AI Act which many believe is damaging the region’s ability to advance in this new sector given that most AI capabilities are being developed outside of Europe. Similarly, there are recent announcements from Apple and Facebook, both have delayed the launch of new AI capabilities on their platforms in Europe due to regulation and policies in the region. More recently, a group of technology companies, including Meta, Ericsson, and Spotify, wrote an open letter to EU Policy Makers. The letter urged the EU to address what they described as “inconsistent regulatory decision making” on artificial intelligence (AI) arguing that these inconsistencies are making Europe less competitive and innovative compared to other regions.

By contrast, the US (by far the current global leaders in the space with OpenAI, META, Microsoft and Nvidia alone) has adopted a decentralized strategy, focusing on public investment, agency-specific guidelines with most regulation occurring at the state level. Billions are being spent on infrastructure build-out supported by countless large-scale joint ventures, with rapid innovation and occurring throughout every aspect of the AI value chain. Innovation and commercialization is occurring in chip design, computation, hardware, Data Centers, Energy Creation, Energy Management and the mind-blowing applications awaiting our discovery. Most of this innovation is occurring in the US right now. ?

Similarly, China (aiming for global AI leadership by 2030), has prioritized rapid development with significant state investment. Billions are being spent to catch-up to the US with little being reported on regulations or policy.

In sum, we have the EU's stringent regulations which aim to mitigate risks and protect the privacy and sovereignty of their citizens, while the US and China focus more on innovation and economic growth. This crudely describes two distinct strategies across a spectrum ranging from Market focus Vs Policy focus.

Canada has an opportunity to take the lead across this entire spectrum, but we have not yet made the regulatory or economic strides required to do so. We have yet to have the pillars in place to reach our full potential in AI – not economically or technologically. However, Canada has an opportunity to take a balanced approach to its AI Strategy. Participating in both the creation of AI capabilities and reaping the economic and social benefits of well-regulated consumption of responsible AI.

Many could argue that Canada has work to do when it comes to its AI mindset, and trust in AI. We could be thinking bigger as a country[BM1]?[NF2]? and there is no assurances about the consequences of falling further behind in this new economy.

?“Canada has an opportunity to take a balanced approach to its AI Strategy”

Describing the Opportunity for Canada | Part 1: Canadian Legacy in AI

As producers of AI capabilities, Canada stands at a pivotal moment in its AI journey and needs to evolve with an approach that couples a strong regulatory foundation with development and innovation. In doing so Canada has a unique opportunity to lead in the AI-driven global economy. By leveraging Canada’s strong research community, diverse talent pool, and robust infrastructure potential in partnership with industry and academia, we can position ourselves as a hub for AI data-privacy and innovation.

Canadians and Canadian-based research are already well recognized for pioneering across the realm of AI, making significant contributions to AI’s creation and propagation. Some of the biggest thinkers defining the parameters of AI are Canadian. So, from a technical context, could we be doing more as a country to augment their work and be a leader in the architecture of AI and how AI is delivered to the world?

Consider such pundits like Yoshua Bengio, of the Université de Montréal (and founder of Mila, Quebec’s Artificial Intelligence Institute); Geoffrey Hinton of the University of Toronto (recent Nobel Prize recipient) and Yann LeCun, a Frenchman who is most groundbreaking research was done at Bell Labs and U of T. These three figures are widely considered to be the godfathers of AI and continue to significantly influence how formative American companies, including OpenAI, Alphabet, Meta, and Anthropic, have established themselves.

While protectionist measures in some sectors in Canada aim to safeguard local interests and outdated thinking, there remains a need for a balanced approach that embraces innovation while ensuring ethical standards and Canada’s data sovereignty.

Describing the Opportunity for Canada | Part 2: Canadian Economy & Productivity

Then there is the topic of Canada’s apparent slowdown in economic growth in recent years. According to a recent article in the CBC titled, Canada is getting poorer when compared to its wealthy peers, data shows, GDP growth lagged population growth in 2023 and we are falling behind our US peers as it relates to GDP per capita. According to this article, US workers are 38% more productive on average as compared to Canadian employees, citing that big firms in Canada spend less on technology (including AI) and upgrading workers’ skills.

?

Unpacking the AI Value Chain (simplified)

To effectively establish laws & regulations to both govern and truly benefit from AI / GenAI, we must first understand the basics about how this technology is created and deployed. Answer questions like, “What is its value chain from the physical world (where the computers live) to the digital universe (where Netflix, Chat GPT and other apps live)”?

To achieve this, we can draw parallels to the OSI framework used to understand and design networking technology by organizing technology layers from hardware (physical) to applications (virtual). By reference, the AI stack ranges from the Hardware layer to the Cloud layer, followed by AI Models, then the Tooling layer, on through to end user Applications. Let us unpack it!

• Hardware Layer: This includes GPUs, CPUs, and TPUs (aka chips) from vendors like Nvidia and AMD, which provide the high-density computational power in computers created by vendors like HP and Dell needed for AI workloads. This hardware is housed in data centres either owned and operated by private companies, governments and/or Cloud-On Ramp companies like Equinix. ???
• Cloud Layer: Cloud platforms like AWS, Azure, and Google Cloud offer secure storage, data connectivity, and AI development tools delivered as software to companies and consumers. ?
• AI Foundational Models Layer: Platforms like OpenAI and MetaAI provide access to Large Language Models (LLMs) and other AI tools.
• Tooling Layer: This includes databases, training tools, and orchestration tools, often hosted in specialized clouds and on specialized infrastructure to monitor, manage, tune, and operate AI.
• Application Layer: The top layer involves AI interacting with users through applications like ChatGPT or CoPilot. Generative AI and Large Language Models can create original content by learning from vast datasets.

Examining this model, two aspects stand-out as areas of focus in support of Canada’s National AI Strategy:

Energy and Electricity:

First, its should be noted that Data Centers consume a lot of electricity and AI workloads, consume on average 10x the among of energy than typical Could workloads. In the creation of a well regulated, responsible AI strategy in Canada, power-management and sustainability are imperatives and would be better supported by a National Energy Strategy that would aim to expand capacity and ensure equitable power distribution across key economic regions and sectors.

Data Sovereignty:

The second aspect to note is the importance of data in all layers of the AI technology stack. AI models run on unexplainably large amounts of data. Without large amounts of data, AI models do not evolve. As the models are fed data, they become more intelligent within the context of the data it is given. The data used to create these models influence the way the models ingest, process, and generate information. In the same way that steps are taken to ensure these models are free of dangerous bias (which involves where the data comes from and how its trained), Canada’s National AI Strategy needs to also account for Canadian content. This would be applicable to AI models being used to create new tool and the tools and apps being consumed by Canadians. To do so, Leaders need to think about Canada’s participation in both the consumption of AI in addition to Canada’s role in the creation and training of AI models being used in Canada, no differently that it would ensuring that there is always a base of Canadian content in media via the CRTC. Our modern day Canadian Content moment is Data Sovereignty in Artificial Intelligence.

“Canada’s National AI Strategy needs to also account for Canadian content.”

The Role of Public-Private Partnerships in Canada’s AI Value Chain (A Spotlight on Energy)

The Canadian government and policy makers need to consider how the AI value chain is brought to life, in Canada and for Canadians. To work through challenges like energy capacity, energy distribution and data sovereignty cannot happen with public resources alone.

On the topic of energy distribution, the vast amounts of electricity used in data centers typically originates from local utility providers who, with the help of government play a key role in deciding how energy distribution is prioritized across sectors and industries. Specifically, AI workloads consume significant amounts of energy. AI workloads also require higher density computing generating vast amount of heat (heat is not good for computers), which in turn require more power to cool them.

Data Centers (DCs) are at the heart of this AI moment: From the DCs that run the “Cloud”, to the DCs that create the AI models themselves, to the DCs that manage the networks that distribute the application we all use. DCs are indeed critical infrastructure.

To remain relevant in the AI realm, the government and utility providers need to be working in tandem with the private sector to reduce barriers to power-delivery and power-consumption to Data Centers. Canada’s AI strategy needs all level of governments support to clear the way for greater domestic delivery, distribution, and access to electricity so that innovation continues to prosper. Today, Canada lacks a national energy strategy and therefore many of such decisions are made at the provincial level and in some cases, AI infrastructure build-out is far from being prioritized.

While there is a view that prioritizing power allocation for critical service such as healthcare, defense, and social services acts as a means of protecting our natural resources and serving our Canadian citizens, we should strongly consider that ensuring electricity distribution is also prioritized to support of a broader AI strategy is also a characteristic of critical infrastructure.

The same aforementioned critical industries (health care, defense, and social services) are all highly dependent on technology services and subsequently, AI. Based on this, considerations should be made to better support the technology that enables many aspects of these critical services which in large part involves investments in Digital Infrastructure and Data Centers.

Simply put, the provision of space and power to eco-friendly data centers and the companies that create and manage this infrastructure is indeed in-service to Canada’s economy, its communities, and its citizens.

In Canada, Data Centers represent 1% of the total electricity used in Canada. That number jumps to 2% in the European Union. 3.5% in the US and 4.2% in China (and poised to grow on average by approximately 38% in all three regions). The distribution of electricity to Data Centers in the US, EU and China will have a direct correlation to those regions success in Digital and AI Capability. In this instance, Canada ranks last among the regions cited.

Knowing that Infrastructure investment would be at the core of any AI Strategy, in doing so Utility Providers, Energy Ministers and Policy Makers need to address energy capacity, energy distribution while considering the downstream impact of those decisions on data sovereignty. ??

There are also innovations in the private sector in power management coupled with avant-garde sustainability strategies that would well serve Canada’s AI Strategy. By teaming and seeking partnerships with innovative organizations with expertise in such areas, government and utility providers have an opportunity to leverage the expertise that currently exist in the private sector to help solve some of the challenges that have led to the suggestion of outdated policies in some Canadian regions.

Strong partnerships between government bodies, energy policy makers, utility providers and data center companies are an essential aspect of Canada’s artificial intelligence strategy – not only for housing AI infrastructure and promoting sovereign AI innovation, but also for doing so with environmental and long-term sustainability in mind.

Recap and Conclusion…

As a matter of economic competitiveness in addition to social and environmental global leadership, Canadian Industry and Government can and need to do more to secure Canada’s advantage in Artificial Intelligence.

Other countries are tackling mid-range challenges and opportunities in the AI space like, the impact of California’s recently vetoed SB 1047 AI Safety bill, (given that most AI technology is being innovated in California with 32 of the top 50 Gen-AI companies headquartered in California. Innovative countries are following the implications of copyright lawsuits on AI training models. Billion-Dollar joint-venture (JVs) are sprouting up weekly (see our recent JV announcements from Equinix & Microsoft), all focused on infrastructure build-out.

Meanwhile, despite having been at the dawn of this AI revolution, Canada is falling behind.

A former Prime Minister once said, “The past is to be respected and acknowledged, but not to be worshipped. It is our future in which we will find our greatness.” Partisanship aside, this sentiment is a reminder of the need for Canada to look forward and embrace the transformative potential of Artificial Intelligence and Gen-AI across its entire supply-chain.

I challenge Canadian lawmakers, to accelerate decision making and get past the foundational topics like power distribution and infrastructure investment that are slowing Canada’s progress in developing our AI value chain. In parallel, I extend an invitation to all parts of government and members of the academic community to partner with the private sector (including companies like Equinix) to shape our National AI Strategy.

Marc Mondésir

Managing Director, Equinix Canada

PS: We didn’t really unpack how Public / Private Partnerships could help tackle the challenges that lay ahead with respect to Data Sovereignty and Canadian Content. We’ll have to tackle that in the next article. Until then, thanks for reading.

Thanks to my writing partners, Sanj, Natasha, Andrew

Collaborators: Sanjeevan Srikrishnan , Natasha Ferraro Andrew Eppich

???????